Faster computers need smaller circuits, like the ones that are a result of advanced research at UCD, reports Dick Ahlstrom.
Computer companies have had great success in making computers faster by making them smaller. Over the past decade the size of individual components built into silicon chips have shrunk to less than one-hundredth the diameter of a human hair.
The researchers making these gains however are finding it tougher to achieve continued reductions in the size of components. We are approaching the point where these components are being measured by the number of atoms across and existing production technologies just aren't up to these demands.
Researchers at University College Dublin's Spectroscopy Research Group hope to change all this by employing a different kind of "light" in the lithographic processes used to burn components into silicon. It is based on producing electromagnetic radiation in the extreme ultraviolet/soft X-ray range where wavelengths are very small.
"The smaller the wavelength used, the smaller the structure you can make," explains Dr Padraig Dunne, senior lecturer in UCD's department of experimental physics. "This drive to higher processing speeds is mainly based on getting smaller structures."
Dr Dunne is a member of the spectroscopy group with team leader Prof Gerry O'Sullivan and Dr Emma Sokell, a lecturer in the department. O'Sullivan and the group recently won a Science Foundation Ireland Investigator Programme award worth €1 million to develop this new extreme ultraviolet (EUV) source that is based on the production of a laser plasma.
UCD scientists were in the laser game early, says Dunne. "I think we bought the first pulsed laser in Ireland back in 1973," he says. And lasers are useful for producing plasmas, the fourth state of matter where energy knocks electrons off to produce ions.
The group specialises in the study of laser plasmas and have expertise in the laser source, the nature of the plasma and the spectroscopic signal from the plasma. Studying spectra, the signal that comes from the interaction between electromagnetic radiation - at its various wavelengths - and matter had been an art in decline. Yet striking targets with lasers to study the spectra emitted by the resultant plasmas still provides valuable data, says Dunne.
"Plasmas are sources of ions and the spectra of a lot of these ions is still unknown," he says. Cataloguing these signals has again become very valuable given the number of new satellites being launched that read molecular and ionic spectra in the EUV/soft X-ray ranges.
"What we noticed with these plasma sources is they are very good at releasing radiation at the EUV and soft X-ray end of the spectrum."
And these have very small wavelengths of around 13.5 nanometres (nm, billionths of a metre). Being able to produce and control an EUV source provides immediate advantages in chip manufacture. "The semiconductor processing industry uses discharge lamps, something like a street lamp with a quartz bulb, to get UV light out."
The UV in turn reacts with chemicals on the silicon surface to burn structures into the chip.
Yet existing technology uses UV with a wavelength of about 193nm. The UCD group's wavelength target is more than 10 times smaller than this and they believe they might be able to build structures in silicon just five nm across.
Dunne believes the group has a real advantage given its longstanding experience with laser plasmas. "We are one of the few specialist spectroscopic groups in the world involved in this work," he says. "We can understand what is going on in the target and in the plasma. We know the basic physics of the atoms and the ions."
The plan is to strike a selected target with a laser, producing a tiny plasma fireball 100 microns across. This in turn will emit EUV radiation, which the team will optimise for brightness.
The EUV in turn will be directed by a group of precisely positioned mirrors that will focus the EUV on the masked silicon ready for lithographic etching, radiation at a minute wavelength for a minute structure. If the team succeeds they will have groundbreaking technology of tremendous value to the world's chip manufacturers.